Automatic Landmark Registration for Bronchoscopy Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current bronchoscopy procedures rely on semi-automated segmentation and registration processes, which are time-consuming and prone to human error, especially in navigating through the complex airway structures of the lungs, where accurate and precise landmark selection is required for effective navigation and treatment.
Innovation Solution
The method involves automatic landmark registration using trajectory information and shape sensing during bronchoscopy, where a catheter navigates through the lung airways, generating location data points and bounding volumes to detect landmarks, and an energy model is created to minimize energy values between detected landmarks, enabling more accurate and precise registration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If semi-automated segmentation and registration processes are used, then manual control and flexibility are maintained, but the process becomes time-consuming and prone to human error
Solution Approach 1:
The system performs automatic landmark detection and registration without requiring manual intervention. The catheter autonomously navigates through airways while the system automatically detects landmarks, generates bounding volumes, and completes registration processes, eliminating the need for operator-performed segmentation and registration tasks.
Solution Approach 2:
The system pre-generates a 3D map of the airways and pre-identifies potential landmark locations before the actual bronchoscopy procedure begins. This preliminary preparation allows the automated system to quickly perform registration during the procedure without time-consuming manual operations.
2Measurement precision
If manual landmark selection is used, then operator judgment is applied, but accuracy and precision in navigating complex airway structures are reduced
Solution Approach 1:
The system replaces manual mechanical landmark selection with an automated computer vision and image processing system. The automated system uses algorithms to detect landmarks, generate bounding volumes from point clouds, and identify centerlines, substituting human operator actions with automated computational processes that provide higher precision.
Solution Approach 2:
The system creates a digital 3D copy of the airway structures from pre-operative CT images and generates a virtual 3D map. This digital replica allows automated landmark detection and registration without requiring direct manual manipulation of physical anatomical structures, enabling precise navigation through complex airway geometries.
3Reliability
If automated landmark detection is implemented, then registration accuracy is improved, but the computational complexity and processing requirements increase
Solution Approach 1:
The automated registration process is divided into distinct sequential segments: (1) generating a 3D map from CT images, (2) detecting landmarks on the 3D map, (3) navigating the catheter and generating point clouds, (4) generating bounding volumes from point clouds, (5) detecting centerlines from bounding volumes, and (6) matching landmarks between the two centerlines. This segmentation allows each computational task to be optimized independently and processed efficiently.
Solution Approach 2:
The system performs computationally intensive tasks in advance, such as generating the 3D airway map and identifying potential landmarks from pre-operative CT images before the actual procedure. This preliminary computation reduces the real-time processing burden during bronchoscopy, allowing the system to focus on tracking and matching tasks during the procedure.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method and system for automatic landmark registration and registration using trajectory information and shape sensing during an endoscopic procedure, such as bronchoscopy, are described herein. A segmentation centerline of airways of a lung may be generated based on a pre-operative computed tomography (CT) image of the lung. Landmarks may be automatically detected on the segmentation centerline corresponding to bifurcations in the airways of the lung. A location data point cloud of locations of a catheter through the airways of the lung during navigation may be generated. A bounding volume of the airways of the lung may be generated and a bounding volume centerline may be detected. Landmarks may be detected on the bounding volume centerline for the same bifurcations. Then, the two sets of landmarks may be mapped as part of registration. The trajectory information with shape sensing may be used to provide non-rigid or fine registration.